To better reject junk and detect overflows. There were already
additional precautions and protections in place, but this is cleaner.
Extended the 1614 unit tests with some more bad syntax cases.
Closes#19475
Also:
- replace `manpage` with `man page`, add to `badwords.txt`.
- badwords.pl: import `-w` feature from curl-www, syncing the two
scripts fully.
- badwords.txt: import missing items from curl-www, syncing the two
files fully.
- pyspelling.words: drop `cURL` allowed word.
Closes#19468
On Windows this requires Perl 5.22 from year 2015.
Also:
- mdlinkcheck: delete redundant error handling logic.
Follow-up to 77be4a7ab2#19437Closes#19461
dynconvert() now offers to return the size of the converted data as it
might be different that the provided input size.
Bonus: minor indent fixing of some closing braces.
Reported-by: Stanislav Fort (Aisle Research)
Closes#19465
The CURL_FTP_HTTPSTYLE_HEAD logic was added back in 2007 with the
intention to remove that logic one day, but since we never bump the
SONAME it is not likely to happen anytime soon. Remove again for
readability.
Follow-up to 3217809294Closes#19463
Verify that wolfSSL_BIO_meth_new() actually works and handle situations
where it returns NULL.
Reported-by: Stanislav Fort (Aisle Research)
Closes#19459
Without unit, show up to 99999 "raw" (5 digits). After that, prefer to
show the number as less than 1000 per unit and use single decimal
fraction. Like '123.4M' (spending 6 characters).
This now makes the largest possible size to show 8.0E (exabytes).
Probably makes the output easier to read.
Fixes#19431
Reported-by: Fd929c2CE5fA on github
Closes#19433
When asking for the last N bytes of a file, and that size was larger
than the file size, it would miss the first byte due to a logic error.
The fixed range parser is now made a common function in the file now
renamed to vssh.c (from curl_path.c) - used by both backends.
Unit test 2605 verifies the parser.
Reported-by: Stanislav Fort (Aisle Research)
Closes#19460
- allow 02_28 to run in HTTP/1.1 without H2 support again.
Follow-up to 3752de465d#19412
- fix to skip 02_28 for all protocols for curl without compression
support (either zlib, brotli or ztsd).
Closes#19458
Also:
- OS400/makefile.sh: use end-of-options marker in xargs command.
- OS400/make-tests.sh: drop warning suppression.
Seems to not trigger anymore as of shellcheck 0.11.0
Closes#19451
Fix SC2046: "Quote this to prevent word splitting."
Ref: https://www.shellcheck.net/wiki/SC2046
Also:
- shellcheck.sh: add `set -eu`.
- shellcheck.sh, yamlcheck.sh: always run from repo root.
- pass `--` before passing the list of files, where missing.
- badwords.pl, cleancmd.pl: rework to accept `git ls-files` arguments.
Requires Perl 5.22+ (2015-Jun-01) on Windows.
Ref: https://perldoc.perl.org/functions/open
- INTERNALS.md: require Perl 5.22 on Windows.
- spacecheck.pl: formatting.
- GHA/http3-linux: rework command to avoid SC2046.
- stop suppressing SC2046 warnings.
The yamlcheck.sh issue reported-by: Stanislav Fort (Aisle Research)
Ref: 20251109163515_6eb31da3-deb2-4f4d-8327-935904f27da5
Closes#19432
Data files no longer depend on mixed newline styles. Before this
patch the harness still assumed data files to use LF newlines,
ensured by `.gitattribute` and distributing sources with LF newlines.
To allow using platform native newlines (CRLF on Windows typically),
update the test harness to support data files with any newline style
on disk. And delete `.gitattributes`.
Fix to:
- load original data files (from test/data) so that their newline-style
doesn't matter on the checked out source repo, meaning it works
when its CRLF on Windows, just like any other file.
(if a BOM slips in, it's caught by `spacecheck.pl` as binary content.)
- do the same in `util.py` used by `smbserver.py` (for test 1451).
- also fix `util.py` to use us-ascii encoding for data files, replacing utf-8.
Also:
- runtests: rework the stray CR checker to allow full CRLF data files,
and keep warning for mixed newlines.
Follow-up to 904e7ecb66#19347Closes#19398
When a blob option is used and it does not convert, the code would
erroneously pass along an uninitialized stack struct.
Reported-by: Stanislav Fort (Aisle Research)
Closes#19444
They make git branch names using those parentheses, that need to be
quoted when used with git command lines. Avoid parentheses for easier to
use branch names.
Follow-up to f77c574445Closes#19441
In the logic called for curl_multi_fdset().
File descriptors larger than FD_SETSIZE size are simply ignored, which
of course will make things break but at least it does not trash memory.
Reported-by: Stanislav Fort (Aisle Research)
Closes#19439
To debug the Windows CI fails further. Acting on the suspicions that
`taskkill` may sometimes be applied to the wrong process.
- log task info, and task child info before calling `taskkill` on a PID.
(on native Windows.)
One of the calls needs PowerShell.
- add env `CURL_TEST_NO_TASKKILL` to disable using `taskkill`.
- add env `CURL_TEST_NO_TASKKILL_TREE` to use `taskkill` without
`-t`, meaning to kill the process, but not child processes.
- GHA/windows: disable `taskkill` calls, to see what happens.
I'll revert or tweak this in a future commit depending on results.
Ref: https://github.com/curl/curl/discussions/14854#discussioncomment-13062859
Ref: https://github.com/curl/curl/discussions/14854#discussioncomment-14913014Closes#19421
Before this patch servers were loading the original data source file
(from `tests/data/test*`) if they failed to open the preprocessed data
file.
It was causing issues in many (most?) tests, because original data files
are not preprocessed, thus may be incomplete and/or come with wrong
newline characters. It's also causing difficult to diagnose issues when
a test accidentally references another test's data, which by chance
makes the test pass initially, until either that or the executed test
data gets an update, and breaking it, as seen in #19329.
Historically, the fallback existed first, then the preprocessed copy.
The fallback is no longer used by tests (except by stray accidents).
Fix it by dropping the fallback logic and relying on the preprocessed
data file saved there by the runtests framework.
Also fix two remaining test data cross-references:
- test1565: reference own server input data instead of test1's.
- test3014: reference own server input data instead of test1439's.
Ref: #19398
Follow-up to aaf9522a2c#19329Closes#19429
If CMake reports the target platform not supporting shared libs, turn
`BUILD_SHARED_LIBS` off by default. CMake 3.30+ fails with an error
when trying to create a `SHARED` target for such platforms. Earlier
versions used a workaround that may or may not have worked in practice.
Ref: https://cmake.org/cmake/help/v3.30/policy/CMP0164.html
Seen this with a build setting `-DCMAKE_SYSTEM_NAME=Generic`, e.g.
AmigaOS.
Note this may introduce incompatibility for "Generic" targets, which
support shared libs. If that's the case, set `BUILD_SHARED_LIBS=ON`
manually.
Also drop AmigaOS-specific logic handled automatically after this patch.
Ref: https://cmake.org/cmake/help/v3.7/command/get_property.html
Ref: https://cmake.org/cmake/help/v3.7/prop_gbl/TARGET_SUPPORTS_SHARED_LIBS.htmlCloses#19420
