Git log: commit 2b5de62086ffab8aa95a248acfa4b0927be8d77d
Author: Yadd <yadd@debian.org>
Date: Tue Apr 16 10:34:56 2024 +0400
Update French translation
commit cdc555e7214403a52beee2dda9a4fbd071acdebd
Author: Agathe Porte <gagath@debian.org>
Date: Wed Mar 20 10:50:12 2024 +0100
uscan: add Fossil example
The Fossil SCM requires non-trivial uscan rules to deduce the tarball
url from the taglist. Adding an example in the uscan man page will help
other people create watch files for such cases.
commit 0953fc7e9dd27c08fc5a656a6a1b4fd63a006d9a
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sat Jan 20 23:55:21 2024 +0100
annotate-output: include prefix format in the annotated output
In order to parse the annotated output it would be nice if that would contain
information about the prefix (format string) it uses.
This is done by printing another line similar to the “header”, which gives the
version information, thereby again altering the output format.
Tools, that parse the annotated output should in principle assume that any lines
`I: …` without a prefix may be included in the output (in principle at any
point).
In order to no need a shell variable (which might unintentionally get passed on
to the actually executed program, see commit
0d5ca640d7261a5554f9918f5ff2a66bb4c0c542 (“annotate-output: prevent `PROGNAME`-
variable from leaking to command”)), the construct with a definition function is
used again.
The double-escaping using GNU `printf`’s `%q`-conversion-specifier is necessary
because once it’s needed for the `eval` and once to have it escaped as output.
The leading `+`, if any, is stripped off because it’s not really part of the
prefix format string, but only needed for the `date`-utility.
The `get_prefix()`-function could have been defined to always print a trailing
`\n`, thereby avoiding the need for the final `printf '\n'`, but the, the
function – should it ever be used for some further purpose – would not return
the real prefix.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit ea10202fa136a76f4c662985e273a4fb8ef26e53
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sat Jan 20 23:46:02 2024 +0100
annotate-output: include version information in the annotated output
As the format of `annotate-output`’s output may change over time, it would be
nice to be able to find out which specific output format was used for a given
transcript (which may be very old).
This commit does so by adding a initial line like:
```
I: annotate-output <version>
```
where the version is set during build via the `###VERSION###`.
An alternative would have been to used fixed output format versions, where one
could now start for example with `2` (assuming `1` was the format which had no
version information).
The problem with that is, that for any future change to the format it must be
remembered to increment the version or otherwise it will be completely useless.
Using the program version however enables one to go through git an see which
exact format was produced by some given version.
The downside is of course, that the version will change with every released
version of `devscripts`.
The initial line does not include a time stamp on purpose, in order to make
parsing it easier. In principle, `file` could now even be extended by some magic
cookie for recognising `annotate-output` files.
There is however a `I: ` prefix. While this could have been omitted it may make
processing of the file easier.
It should be noted, that this commit changes the output format of the program.
The benefits should however outweigh this.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit a772ce8be24d289444f4b0afcc365ccb61b9989b
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sat Jan 20 23:41:56 2024 +0100
annotate-output: use single-quotes where possible
Use double-quotes only where they are needed.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 4c761e3a1b73340108b37f714e3ddf1c45d84250
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sat Jan 20 20:25:44 2024 +0100
annotate-output: escape command name and arguments in the annotated output
Previously, an invocation like:
```
$ annotate-output -- echo a 'b c'
```
would have caused an annotated output like:
```
20:26:49 I: Started echo a b c
20:26:49 O: a b c
20:26:49 I: Finished with exitcode 0
```
where from the `I: Started …` line it would have no longer been possible to
determine the true arguments.
Similarly, an invocation like:
```
annotate-output.sh -- echo a $'b\nc'
```
would have caused an annotated output like:
```
20:28:30 I: Started echo a b
20:28:30 I: c
20:28:30 O: a b
20:28:30 O: c
20:28:30 I: Finished with exitcode 0
```
which would have no longer followed the expected format (and thus be not easy to
parse) and where it would have no longer been possible to determine the true
arguments.
This commit, escapes the command name and arguments using GNU `printf`’s `%q`-
conversion-specifier.
Now the above invocations will produce the output:
```
20:26:49 I: Started echo a 'b c'
20:26:49 O: a b c
20:26:49 I: Finished with exitcode 0
```
respectively:
```
20:28:30 I: Started echo a 'b'$'\n''c'
20:28:30 O: a b
20:28:30 O: c
20:28:30 I: Finished with exitcode 0
```
For the rationale of using GNU `printf`, see commit
bef65a9fcd68ad0a585edac6e969a093092f8c8f (“annotate-output: prevent `FMT`-
variable from leaking to command”).
Since that line is printed only once, the additional cost of invoking a non-
built-in `printf` should be negligible.
It should be noted, that this commit changes the output format of the program in
the sense that when invoked exactly the same way, the “Start …” line may still
look different due to the escaping.
The benefits should however outweigh this.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 5cb24fead3ebcfd4129d603fb0f78e4a83802226
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Mon Dec 11 01:50:11 2023 +0100
annotate-output: describe caveats on environment variables
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 95b20ae7e1a34db7b4a9f59933d4bca361d912dc
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Mon Dec 11 01:26:13 2023 +0100
annotate-output: improve documentation of bugs/caveats
Describe more clearly which output (to STDOUT and STDERR) is expected from the
invoked program.
Especially notice that NUL bytes always lead to undefined behaviour as POSIX
shells cannot read them (and it’s undefined what they do, when they encounter
one).
Other bytes may lead to undefined behaviour, depending on the locale that’s
being used. For `dash` this is however not a problem, as that doesn’t support
locales other than the `C`/`POSIX` locale in which all bytes are valid
characters (other than NUL).
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 1be19b890a9a80605be70a7b2fa5a3bdbf29408f
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Mon Dec 11 01:10:12 2023 +0100
annotate-output: reformat program usage
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 719db105af9b8d28816b0cfe2cec4f96ffab9873
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Mon Dec 11 01:00:48 2023 +0100
annotate-output: unify spelling of `STDOUT` and `STDERR`
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit bef65a9fcd68ad0a585edac6e969a093092f8c8f
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 22:50:40 2023 +0100
annotate-output: prevent `FMT`-variable from leaking to command
Analogous to commit 0d5ca640d7261a5554f9918f5ff2a66bb4c0c542 (“annotate-output:
prevent `PROGNAME`-variable from leaking to command”).
This removes the need for the `FMT`-variable altogether, by “hardcoding” it into
the new generically (via `eval`) performed definitions of the
`handler`-function.
First, the function is always defined once to get the default of `+%H:%M:%S `
set. Then, depending on the program options, the function may be re-defined
numerous times (this was also already the case, when there were not even the
`define_handler_*`-functions and is also documented meanwhile.
Using `eval` is always tricky:
- The leading space in `eval " handler() {` is not strictly necessary here, but
nevertheless good practise, as `eval` is not required to support `--` and
POSIX recommends a leading space instead.
While we don't evaluate a string with a leading `-`, this might in principle
change in the future.
- As this commit wants to “hardcode” the user-supplied format string into the
function it must be escaped, otherwise unusual format strings like:
```
$ annotate-output '+foo"; ls; printf "bar' date
```
(or smilar, depending on the actual non-escaping implementation)
would lead to code-injection (above, the `ls` would be executed).
Properly escaping a string for literal re-use as shell input is difficult, but
some implementations of `printf` have added the – as of now still – non-POISX
`%q` conversion specifier which performs the escaping.
I myself has requested this to be standardised but was too late for the
upcoming new revision of POSIX, though the working group agreed to at least
reserve `%q` and `%Q` for that purpose (see
https://austingroupbugs.net/view.php?id=1771).
So it’s actually not so unlikely, that this will become true POSIX sooner or
later.
Anyway, meanwhile, `printf` from the GNU Core Utilities can be used.
It’s anyway executed only at most twice per invocation and since the
`coreutils` is an essential package in Debian, no new dependency is intrduced
by that.
With this commit, there should be no more variables set by our script that may
be exported into the invoked program.
Only those environment variables for which the shell (usually `dash`) sets its
own variables unconditionally or that are lost because they’re not valid shell
variables, may not be passed on properly from a caller to the invoked program.
This would however require a re-implementation in a language like C.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit e4d788066a60901d8ad9b2c36e652c0ea2761211
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 19:59:52 2023 +0100
annotate-output: sanitise `IFS` when reading the exit status
Again, strictly speaking not necessary, but costs nothing and indicates that
whatever `IFS` might ever be changed to in the future within the script, the
read must not do field splitting.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit e3ef1c1953285aed7cadc73877ced7b86917ef46
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 19:40:24 2023 +0100
annotate-output: switched from `echo` to `printf`
`echo` cannot portably be used, unless `-n` is used and escape sequences are
omitted.
Especially the `$0` and `$@` could have included escape sequences.
Since `printf` is anyway also built-in in all reasonable shells, switched
everywhere to that.
It does not matter, that `$?` and `$err` are used directly in the format string,
as these can only ever contain digits which will never get special meaning in
there.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 1b7c805df5ea7424a771cccdf6f4bdfa668750e8
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 19:26:29 2023 +0100
annotate-output: support `--` to end option parsing
Previously, it would not have been possible to execute a program that has the
name of an option recognised by `annotate-output`.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 0791eeb6376f51eec633f5d19b81846b64b5ddd4
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 19:05:45 2023 +0100
annotate-output: added option that allows date format string with no separator
This option mainly allows to print the date format string as `+FORMAT` does, but
without the following space that separates it from the stream indicator, which
may be of use if one wants to print no date at all (what currently causes a
useless leading space).
It’s added as a new option, to leave `+FORMAT` backwards compatible.
The name `--raw-date-format` was chosen, rather than for example just
`--raw-format` as future versions of the program may want to allow for more
generic format strings (for example such, where the stream indicator or even the
line is also a conversion specifier).
If no format is specified, the program exits with `125`.
This is used for the same purpose by several GNU Core Utilities (for example
`nice`, `nohup`, `stdbuf` and `timeout`).
POSIX.1-2017 reserves in principle only `126` and `127` as “special” exit
statuses and both do not really fit (though it does indeed specify the `127`
exit status for the `nohup` utility to mean both, the utility itself having
failed or the utility invoked by it having not been found).
It does however specify the exit statuses `1` up to `125` for the `nice` and
`time` utilities to mean an error in the respective utilities themselves, which
overlaps with exit statuses that may be returned by the utilities invoked by
them.
Since `annotate-output` is conceptually similar to `time` (annotating output) it
was chosen that `125` is a proper choice, despite not being 100% POSIX.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 45e52639db4e0ff41df2b5574b7b83d3fc0de4cb
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 18:12:22 2023 +0100
annotate-output: print concrete error message rather than just the usage
If the actual problem is known, it’s better to print a concrete error message
than the rather useless program usage, which doesn’t tell the user what went
wrong.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit ee5ba3d81f87757354ffb3161643f9130bcb98bc
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 18:05:58 2023 +0100
annotate-output: improved program usage output
Align it with the changes to the manpage in the previous commit.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit ed70be9dc93c70d94d21d01346433ebde553d669
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 18:03:37 2023 +0100
annotate-output: improved the manpage
This prepares for changes from follow-up commits by improving the description of
the program output and the format string, in particular by adding missing
information, like the documentation of the `I`-stream indicator and the exact
formatting of the whole lines.
It further applies upper-case spelling of `STDOUT` and `STDERR` like it’s
already used in the program’s usage.
Further it more clearly differentiates between “this” program (that is
`annotate-output`) and the program that is executed by it.
It also indicates that multiple options can in principle be given.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 8eda551704c3645941e2d44ac762e9891d54a3f4
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 17:17:15 2023 +0100
annotate-output: removed undocumented `-help`-option
Long option names with only one leading `-` are rather uncommon and cause
problems should further short options be added.
Since the option is anyway not documented, neither in the manpage nor in the
program’s usage output, it’s simply removed.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 61e4707eac2fc624f09d963497de053f1a907239
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 17:15:54 2023 +0100
annotate-output: documented exit statuses
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 91785e6f94344c754decaf6f70e10f740afb26da
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 16:58:27 2023 +0100
annotate-output: use exit status `127` if no program was specified
If after option parsing no arguments remain, no program to be executed was
specified.
POSIX.1-2017 reserves the exit status `127` for cases when a command to be
executed was not found in several places of the standard:
- For the Shell Command Language in chapters “2.8.2 Exit Status for Commands”
and “2.9.1 Simple Commands”.
- For the `exec`-special-built-in-utility in chapter “2.14. Special Built-In
Utilities”.
- In general, in the rationale of the `exit`-special-built-in-utility in “2.14.
Special Built-In Utilities”
- For the `nice`-, `nohup`-, `sh`, and `time`-utilities in chapter “4.
Utilities”.
(all in “Volume 3, Shell and Utilities”).
This seems to also be the best fit, for the case when no program was specified.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit c78099d4b01e291bdc1229fb542649e5bdcf3e85
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 16:09:19 2023 +0100
annotate-output: double quoted parameter expansions
The script does (and should) not set `IFS` to a sane value, thus in principle it
may contain digits and thereby affect parameter expansions of positional and
special parameters.
While this is in principle not necessary for the assignments, it’s done there
nevertheless for the case that the code would be reused in some places where
field splitting does happen.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 9d0d2c53298a0be4e243c3021d81fb86648f24cc
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 07:36:19 2023 +0100
annotate-output: prevent `err`-variable from leaking to command
Analogous to commit 0d5ca640d7261a5554f9918f5ff2a66bb4c0c542 (“annotate-output:
prevent `PROGNAME`-variable from leaking to command”).
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 6128a453e834c126215859803b54f2f2e7e36c39
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 07:17:42 2023 +0100
annotate-output: removed static separator between date/prefix and stream indicator
This prepare to allow for removing the separator (a space) between the
date/prefix and stream indicator in a future commit.
It should not have any effect on the output format as it is.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 9522e723dc58ca9804a18a53ea5ee0f0c7011f0f
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 07:09:00 2023 +0100
annotate-output: removed no longer needed code
These remnants became obsolete with the previous commit.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit fbb1da47a8aefd0fc3ce5c873cc09b839b465bd1
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 06:42:13 2023 +0100
annotate-output: split up `handler`-function for date and plain cases
This splits up the single `handler`-function into two, one for the case when a
date conversion specifier is used and one for the case when not.
It uses the previously introduced definition function approach, to get the
`handler`-function defined as either of its two implementations – but always
with the same function name.
The respective definition function is called, depending on the format string.
The benefit of all this are performance improvements.
On my system, with a date conversion specifier (the default one):
```
$ time annotate-output.sh tree /usr/share/doc
```
takes previously: against now:
real 1m14,336s real 1m5,457s
user 1m2,140s user 0m52,115s
sys 0m17,953s sys 0m17,058s
The improvements in this case come likely from the no longer needed call to the
`datefmt`-function and perhaps a bit from not needing to expand `$1`.
~10s is admittedly not much, but the gain gets far bigger when not having a
date conversion specifier:
```
$ time annotate-output.sh + tree /usr/share/doc
```
takes previously: against now:
real 0m19,067s real 0m2,418s
user 0m11,738s user 0m0,910s
sys 0m9,784s sys 0m2,585s
The improvements in this case come likely from the no longer needed subshell
(from the command substitution) and perhaps a bit from the saved `printf`.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 566efd76b54d636c48af00e899bbdb5b5713f9f7
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 06:26:48 2023 +0100
annotate-output: define `handler`-function in a definition function
This prepares for subsequent commits and merely moves the definition of the
`handler`-function into a definition function, which is then called.
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit 0d5ca640d7261a5554f9918f5ff2a66bb4c0c542
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 05:42:22 2023 +0100
annotate-output: prevent `PROGNAME`-variable from leaking to command
annotate-output is a wrapper program that runs other programs. Such programs may
in turn make use of arbitrary environment variables.
Every variable that is set within the script, may have been already in the
environment of the shell (because it was already exported when `annotate-output`
itself was called).
The script would then overwrite the variable and export it further to the actual
program, which would not get the value as expected by the user.
A simple example (before this commit) would have been:
```
$ export PROGNAME=foo; annotate-output env | grep -E ^.{12}PROGNAME=
15:50:19 O: PROGNAME=annotate-output
```
Unfortunately, the problem cannot be fully solved in the Shell Command Language,
amongst others because:
- Shell always unconditionally set some own variables - which again my have
already been marked as exported.
- Environment variables need to to be valid shell variables (e.g. `*=value`) is
in principle a valid environment variable, but not a shell variable) and it’s
unspecified whether such are still passed on to commands (called from the
shell) or not.
A re-implementation in e.g. C would be needed to overcome this limitation (side
note: even Python cannot be made 100% transparent of the environment, see
https://github.com/python/cpython/issues/93446).
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit b5c6a21ff1ba0b0de3616fe23d652f503f206be4
Author: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
Date: Sun Dec 10 05:23:36 2023 +0100
annotate-output: fixed typo
Signed-off-by: Christoph Anton Mitterer <mail@christoph.anton.mitterer.name>
commit a323c636546ec56c69133cb15c0bc34db5e5d98a
Author: Johannes Schauer Marin Rodrigues <josch@mister-muffin.de>
Date: Thu Jan 18 17:24:59 2024 +0100
scripts/bts.pl: status: preserve field order and otherwise sort keys
The order of items returned by keys was not deterministic. This means
that both with and without "fields:", the bts status output was having
the lines per stanza semi-randomly ordered. Since it is unlikely that
any application relies on semi-randomly field order, this commit:
- sorts the keys alphanumerically if no "fields:" is given or
- print the keys in the same order as they are given via "fields:"
Applications parsing the output of bts status already had to account for
semi-random field ordering, so they should not break with the output
being deterministic now. If "field:" is not passed, the bug_num field
will remain the first line per stanza. Only the remaining fields are
sorted. This preserves the existing behavior concerning the position of
the bug_num field in the output.
commit ff22a7f075b5ca88bcd50bd1648f0c039f44c435
Author: Michael Prokop <mika@debian.org>
Date: Tue Dec 19 08:57:29 2023 +0100
README: update dependencies for debootsnap + add python3-pycurl to Suggests
debootsnap depends on python3-debian + python3-requests,
though dpkg-dev is already a hard dependency (so drop it).
Add python3-pycurl to Suggests.
Closes: #1054571
commit d69b137efdd883b3adc34401c7689802eda3ca70
Author: Sylvestre Ledru <sylvestre@debian.org>
Date: Fri Feb 2 09:22:56 2024 +0100
find_watch_files: replace the deprecated -follow in find by -L
commit 04d446b9cec4a8c08fbd248dece649a1c6b281d3
Author: cen <imbacen@gmail.com>
Date: Thu Feb 15 15:55:13 2024 +0100
add missing import for field_parse_binary_source, invoke conditionally on whitespace present in package name
Closes: #1063928
commit d0550188cc1634526abc0641af6e4db5648706f6
Author: Paride Legovini <paride@debian.org>
Date: Wed Feb 28 12:13:49 2024 +0100
wrap-and-sort: make help text more uniform
Make the capitalization, punctuation and length of help text more
uniform.
commit bba719ea09e9242dd4b2f78337ab98b265125ec0
Author: Paride Legovini <paride@debian.org>
Date: Wed Feb 28 11:17:23 2024 +0100
wrap-and-sort: hide deprecated --experimental-rts-parser from help
commit ddcb72510408ee92f49c2c1cc15e780bce521b1f
Author: Paride Legovini <paride@debian.org>
Date: Wed Feb 28 11:15:50 2024 +0100
wrap-and-sort: update manpage with the new --no-<action> arguments
commit a3d9e684fe6daef1378488a31841db05db4e3372
Author: Paride Legovini <paride@debian.org>
Date: Wed Feb 28 12:04:28 2024 +0100
wrap-and-sort: show defaults in help using ArgumentDefaultsHelpFormatter
commit ddb6ef11bd9e530bd4df4db1838bb15083eae8ee
Author: Paride Legovini <paride@debian.org>
Date: Wed Feb 28 12:02:35 2024 +0100
wrap-and-sort: consisteny add --no-<action> options
commit 888970f761029977edca5317bf9f27c2bd0a8b73
Author: Otto Kekäläinen <otto@debian.org>
Date: Fri Dec 29 11:36:28 2023 +0800
Improve dget help and man page to clarify difference to modern apt-get
The man page of dget had incorrect references to 'apt-get' (probably due
historical reasons before apt-get got improved) and it was unclear what
new Debian Developers are supposed to use dget for. Clarify the main use
case and add examples to make the utility of dget clean, and also update
the 'apt-get' reference to list the one repository download case apt-get
does not yet handle (get all binaries for source package name).
Related: Bug#1059551
commit 1789a0646f77f68903f20db595adb1343c0981af
Author: Johannes Schauer Marin Rodrigues <josch@debian.org>
Date: Mon Feb 26 13:55:59 2024 +0100
xz output changed from version 5.4 to 5.6. Compare uncompressed data to not depend on the xz-utils version
commit 3d671fde8ef14124442c3fc5089e228c4aa23995
Author: Johannes Schauer Marin Rodrigues <josch@debian.org>
Date: Sun Sep 25 12:08:28 2022 +0200
scripts/debootsnap: always output tarball independent on how the output path is named
commit 827bc068609f1b55866265ca2f9433248f062aa6
Author: Johannes Schauer Marin Rodrigues <josch@mister-muffin.de>
Date: Fri Feb 16 09:15:07 2024 +0100
scripts/reproducible-check: reformat with black 24.2
commit dd958aed7016a02c6e0bf4e9082a7a6ea7c16df2
Author: Johannes Schauer Marin Rodrigues <josch@mister-muffin.de>
Date: Fri Feb 16 08:29:35 2024 +0100
Revert "ci: skip running tests on currently broken debian:testing"
This reverts commit 6f1e960c5b94298f9decd2850da2dbbc66250439.
commit 3325b66eb1a89053b2ea881304aaf603073e89cc
Author: Johannes Schauer Marin Rodrigues <josch@mister-muffin.de>
Date: Tue Feb 6 08:05:12 2024 +0100
ci: skip running tests on currently broken ubuntu:devel
This is similar to 6f1e960c5b94298f9decd2850da2dbbc66250439 which
disabled testing due to python-flake8 issues.
commit 591085e8ede4764a4026bf9285990dad6701d59c
Author: Gianfranco Costamagna <costamagnagianfranco@yahoo.it>
Date: Fri Feb 23 08:59:24 2024 +0100
Fix uscan mode=git to use the correct versioning sha when a branch is provided (Closes: #992976)
commit 35c437c56a4e173c9c3ebf27e0e5fb96a8c2de3c
Author: Yadd <yadd@debian.org>
Date: Fri Feb 2 00:44:54 2024 +0400
Add deb2docker and deb2apptainer scripts. Thanks to Emmanuel FARHI.
commit ad34061d135eb96031a672cc646842ff03d87f7a
Author: Johannes Schauer Marin Rodrigues <josch@debian.org>
Date: Mon Feb 20 06:44:38 2023 +0100
debbisect: pipe dpkg output into cat instead of using --no-pager
The --no-pager option is only supported by dpkg 1.19.2 or newer (October
2018).
commit 414abeb1a4e469c500ef4afd5ab476a6d3c8d868
Author: Johannes Schauer Marin Rodrigues <josch@debian.org>
Date: Mon Oct 30 09:09:50 2023 +0100
scripts/debbisect: also catch BrokenPipeError as otherwise half-downloaded files are not cleaned up in case apt quits the connection
commit 68f50f25a9bbdb172d4bb9ade250a1c7f7aec06c
Author: Yadd <yadd@debian.org>
Date: Tue Jan 9 07:25:58 2024 +0400
Fix broken bash_completion for salsa(1)
commit fe9f3e3cf1866e39b7f27fd71e896a7de6fc1fcc
Author: Yadd <yadd@debian.org>
Date: Thu Dec 28 18:19:31 2023 +0400
Update French translations
commit 3285f59666029737a72671ac0cda6762593a1bb7
Author: Paul Wise <pabs@debian.org>
Date: Wed Oct 18 07:08:55 2023 +0800
build-rdeps: add debugging for apt indextargets filtering process
Useful when build-rdeps is not finding any Packages/Sources and
the reasons why that is happening are not immediately obvious.
commit c75283019f4ae44798a243a253f887e6b7e7421c
Author: Paul Wise <pabs@debian.org>
Date: Wed Oct 18 15:58:38 2023 +0800
build-rdeps: factor out debug printing into a function
Suggested-by: Guillem Jover <guillem@debian.org>
Suggested-in: https://salsa.debian.org/debian/devscripts/-/merge_requests/370#note_435134
commit 731ad46dd6512001d1f5d11d2b247684ffc3c125
Author: Paul Wise <pabs@debian.org>
Date: Wed Oct 18 07:04:47 2023 +0800
build-rdeps: check devel suite status based on chosen origin not dpkg vendor
Allows using APT_CONFIG and an origin override to check Ubuntu chdists.
commit b80680afd5f94c39ca29f3420b3496e9fa5bf457
Author: Paul Wise <pabs@debian.org>
Date: Wed Oct 18 07:02:18 2023 +0800
build-rdeps: set the default origin to the current vendor from dpkg
Avoids Debian derivatives needing to add patches to override it.
commit 6cac1ed74c80c42de7d142320737ea567fb3d9ff
Author: Paul Wise <pabs@debian.org>
Date: Wed Oct 18 15:55:59 2023 +0800
build-rdeps: avoid Perl errors when a source has no Suite or no Codename
The filtering code currently assumes Suite/Codename both exist but
apt sources generated by local-apt-archive do not include them.
commit 5e948fba0cc2a87112b8605d6459af4745b4b552
Author: Osamu Aoki <osamu@debian.org>
Date: Sun Nov 12 14:01:34 2023 +0900
dget: add deb822 *.sources support
- Relax over restrictive sanity check
- Stop parsing APT configuration with regex
- Use $host_re matched $(BASE_URI)
Closes: #976673, #1030277
Signed-off-by: Osamu Aoki <osamu@debian.org>
commit d82ecfee3eecd96cc0bbbf6e28b790d2faf05706
Author: Benjamin Drung <benjamin.drung@canonical.com>
Date: Tue Dec 12 22:37:02 2023 +0100
rename README to README.md
