Age | Commit message (Collapse) | Author | Files | Lines | |
---|---|---|---|---|---|
2013-12-15 | regtest: Add a command line option to create-report command to not launch ↵ | Carlos Garcia Campos | 2 | -3/+7 | |
the browser | |||||
2013-12-15 | regtest: Return an exist status code depending on whether the command succeeded | Carlos Garcia Campos | 8 | -10/+20 | |
2013-12-06 | regtest: Limit the stderr files to ~1MB | Carlos Garcia Campos | 1 | -1/+5 | |
Some tests send a lot of information to stderr, usually due to parsing errors in buggy documents. More than 1MB of stderr output is diffcult to hanlde and in most cases it's redundant with a lot of duplicated messages. This patch reduced the size of the refs dir for the complete test suite by 1GB. | |||||
2013-12-06 | regtest: Save checksum results sorted in md5 files | Carlos Garcia Campos | 1 | -1/+1 | |
We are using os.listdir() to get the list of test results that returns files in arbitrary order. | |||||
2013-12-06 | regtest: Show also the tests expected to crash/fail to run but don't fail | Carlos Garcia Campos | 1 | -1/+7 | |
2013-12-06 | regtest: Improve readability of test results | Carlos Garcia Campos | 1 | -20/+47 | |
Show a summary of tests failed per backend with the percentages and use a new line for every test in the result instead of using a comma separated line. | |||||
2013-11-29 | regtest: Do not buffer stderr output | Carlos Garcia Campos | 1 | -10/+30 | |
Some buggy documents can produce a huge stderr output because of parsing errors or whatever. We could give a file directly to Popen to write the stderr file, but we only want to create the file when there's output, because it's what we use to know whether the command produced output or not. So, instead of buffering the whole output and then write it to the file, now we read from the pipe while the command is running, writing the output in chunks to the file. This improves a lot the memory consumption when running some tests. | |||||
2013-11-29 | regtest: Read test results in chunks to get the md5 digest | Carlos Garcia Campos | 1 | -7/+12 | |
Some backends can generate huge results, like huge postscript files that we don't want to load in memory to get the md5. So, instead of creating thr md5 object with the entire file, we feed it with chunks of data using the update method. This improves a lot the memory consumption and performance as well. | |||||
2013-11-29 | regtest: Remove unused method _check_exit_status2 | Carlos Garcia Campos | 1 | -20/+0 | |
It was used when the backends ran in parallel odd and even pages, but it's no longer used since threads support was added. | |||||
2013-11-29 | regtest: Do not store the current line in Printer but only its length | Carlos Garcia Campos | 1 | -5/+5 | |
We are not using the line text anymore, but only the length. | |||||
2013-11-27 | regtest: Do not consider docs with no refs as skipped | Carlos Garcia Campos | 1 | -17/+23 | |
Handle them differently as new docs, and show them in the summary suggesting to use create-refs command to include them in the test suite. | |||||
2013-11-27 | regtest: Use number of tests run to show the progress | Carlos Garcia Campos | 3 | -112/+118 | |
Instead of using the number of document tested. We don't really need to set a number to a particular document, we only want to know the progress of the whole process. This ensures that the progress is shown in the correct order. It also simplifies the code a bit. Also improved the output formatting to make it easier to read. | |||||
2012-12-29 | regtest: Print test results when the test has finished | Carlos Garcia Campos | 2 | -15/+28 | |
To make sure the result corresponds to the test now that we are using multiple threads. | |||||
2012-12-08 | regtest: Use the number of cpus as default number of worker threads | Carlos Garcia Campos | 1 | -3/+5 | |
2012-12-08 | regtest: Remove trailing whitespaces | Carlos Garcia Campos | 3 | -13/+13 | |
2012-12-06 | implement parallel testing using Python's Queue class | Adam Reichold | 6 | -34/+104 | |
2012-10-21 | regstest: Add print_test_result_ln() and use it for failed test results | Carlos Garcia Campos | 2 | -5/+8 | |
To make sure they are always shown. | |||||
2012-10-21 | regtest: Do not rewrite lines in verbose mode | Carlos Garcia Campos | 1 | -1/+1 | |
2012-09-23 | regtest: Reduce the noise of the default output when running tests | Carlos Garcia Campos | 8 | -29/+137 | |
Show permanent information only about failed tests, without the details about the failing pages. Previous verbose output is available passing --verbose command line output. | |||||
2012-04-01 | regtest: remove debug print to show bisect exit status in Bisect.py | Carlos Garcia Campos | 1 | -1/+0 | |
2012-04-01 | regtest: Fix crash when creating HTML report for a single test | Carlos Garcia Campos | 1 | -4/+1 | |
2012-04-01 | regtest: Don't show results for crashed or failed to run tests in HTML report | Carlos Garcia Campos | 1 | -0/+2 | |
2012-03-11 | regtest: Add find-regression command to run git bisect automatically | Carlos Garcia Campos | 4 | -0/+339 | |
2012-03-10 | regtest: Add information about skipped tests in the summary | Carlos Garcia Campos | 1 | -3/+9 | |
2012-03-03 | regtest: Ignore backends with no results when creating html report | Carlos Garcia Campos | 1 | -0/+5 | |
2012-03-03 | regtest: Add --pretty-diff option to create-report command | Carlos Garcia Campos | 2 | -3/+129 | |
It includes a pretty-diff link to make a bit easier to check differences in test results. For images it creates a html with javascript that toggles actual and expected image, using the same code than WebKit layout test results. For text files it uses HtmlDiff class from python difflib. It's an option disabled by default because pretty diff for text files is very slow. | |||||
2012-02-25 | regtest: Add create-report command to generate html report of test results | Carlos Garcia Campos | 3 | -1/+273 | |
2012-02-25 | regtest: Use diff.png extension for image diff files | Carlos Garcia Campos | 4 | -5/+17 | |
So that they are recognized as images | |||||
2012-01-07 | regtest: read stderr output before calling wait to fix a possible deadlock | Thomas Freitag | 1 | -4/+4 | |
In the way the scripts starts it subprocesses, stderr becomes buffered for the subprocesses. And unforunately, when the buffer limit is reached, the subprocess suspends it work until it can print again on stderr. That's why the python script runs into a deadlock when the subprocess produces a lot of error messages. A small rearrange of the commands, first read the stderr output and then wait that the subprocess ends, will remove this deadlock. | |||||
2011-11-06 | regtest: Skip tests with results when creating refs | Carlos Garcia Campos | 2 | -2/+5 | |
Not only tests that have md5, but also crashed or failed tests. | |||||
2011-11-01 | regtest: Add --update-refs command line option to run-tests command | Carlos Garcia Campos | 3 | -3/+37 | |
It allows to update the references of failing tests. | |||||
2011-09-20 | Update .gitignore | Adrian Johnson | 1 | -0/+1 | |
2011-09-19 | regtest: render cairo at 72ppi | Adrian Johnson | 1 | -2/+2 | |
For consistency with splash. It is faster and easier to compare with splash results. | |||||
2011-09-19 | regtest: Ignore checksums of crashed, failed and stderr files | Carlos Garcia Campos | 1 | -3/+3 | |
They shouldn't have a checksum, but it might happen with md5 files generated with previous versions. | |||||
2011-09-18 | Render at 72 instead of 150, it's faster :D | Albert Astals Cid | 1 | -2/+2 | |
2011-09-18 | regtest: Do not create checksums for crashed, failed and stderr files | Carlos Garcia Campos | 1 | -1/+11 | |
2011-09-18 | regtest: Fix checksum check for tests containing spaces in filename | Carlos Garcia Campos | 1 | -1/+1 | |
2011-09-14 | Make it work with Python3 | Albert Astals Cid | 9 | -38/+38 | |
2011-09-13 | regtest: Add a way to skip files | Carlos Garcia Campos | 4 | -5/+39 | |
A new command line option --skip has been added to give a file with the list of test to skip. When --skip is not used, it look for a file named Skipped in the tests dir. Lines starting with '#' are considered comments and are ignored. | |||||
2011-09-13 | regtest: Limit the number of arguments to 1 | Carlos Garcia Campos | 3 | -26/+26 | |
It's easier to run poppler-regtest more than once if you need to run different tests. | |||||
2011-09-13 | regtest: Sort tests before create-refs/run-tests and show progress | Carlos Garcia Campos | 3 | -23/+53 | |
2011-09-13 | Add initial poppler regressions test program | Carlos Garcia Campos | 14 | -0/+1025 | |