Re: [Libreoffice-qa] writing unit test for calc functions
On 16.4.2016 12:20, Markus Mohrhard wrote: @QA team: Does this sound like something that would make it easier for you to add tests for spreadsheet functions? Hi Markus, both variants are similarly arduous, so it's developer decision. If these new test will lead to easier incorporating of new tests, then +1 from me. In fact it's only way, because we have not devs capacity for csv tests (you noted line endings (windows vs linux), string and number handling, rounding issues on the orcus and calc side, ...) Regards Raal ___ List Name: Libreoffice-qa mailing list Mail address: Libreoffice-qa@lists.freedesktop.org Change settings: https://lists.freedesktop.org/mailman/listinfo/libreoffice-qa Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/ Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette List archive: http://lists.freedesktop.org/archives/libreoffice-qa/
Re: [Libreoffice-qa] writing unit test for calc functions
Hey Pedro, On Sat, Apr 16, 2016 at 5:05 PM, Pedro <pedl...@gmail.com> wrote: > Hi Markus > > Personally I'm very interested in using Calc as a replacement for Excel. > The > spreadsheets I work with commonly have more than 65k lines so using the old > Excel 2003 (included in my work desktop) is useless but Calc is still not > stable/efficient enough. > > One of the main problems with Excel and Calc is compatibility. ODS is out > of > question (because MS manages to ruin all ODS files) and XLSX has lots of > issues in round-trip work (particularly in Pivot Tables) > This whole stuff here is purely about the formula compiler and formula interpreter. For other features exist other types of tests, filters-test, export-test, ... > > Regarding your CSV only approach, Calc does have the option to Save > formulae > instead of values (and curiously opening these files in Excel 2010 works!) > but there are probably limitations (I just found out that array formulas > are > not interpreted correctly on both programs) > The CSV approach is already implemented and is what we are currently using. It is basically just a comparison of the formula results with a csv file. So the question right now is if the csv part should be replaced by a pure spreadsheet error checking. The CSV approach always required a developer as there are many special cases that can not completely be handled without running the test. > > Since it is not in Microsoft's interest to be 100% compatible with ODS, I > think that your spreadsheet approach should include both ODS and XLSX. > These tests are for the formula compiler and formula interpreter. Everything that can be evaluated can be expressed in an ODF file (maybe not visible to the user). > > So, if manual testing your files leads to greater compatibility you can > count me in. > The idea is to generate test files that can be used for manual testing but more importantly can also be used in the automated testing. Currently all our formula compiler and formula interpreter tests can only be used in the automated testing. One advantage of the new approach would be that it makes it easier for non-developers to write such test documents and make therefore sure that all functions are properly tested. > > Regards, > Pedro > > > > -- > View this message in context: > http://nabble.documentfoundation.org/Libreoffice-qa-writing-unit-test-for-calc-functions-tp4181287p4181297.html > Sent from the QA mailing list archive at Nabble.com. > ___ > List Name: Libreoffice-qa mailing list > Mail address: Libreoffice-qa@lists.freedesktop.org > Change settings: > https://lists.freedesktop.org/mailman/listinfo/libreoffice-qa > Problems? > http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/ > Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette > List archive: http://lists.freedesktop.org/archives/libreoffice-qa/ ___ List Name: Libreoffice-qa mailing list Mail address: Libreoffice-qa@lists.freedesktop.org Change settings: https://lists.freedesktop.org/mailman/listinfo/libreoffice-qa Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/ Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette List archive: http://lists.freedesktop.org/archives/libreoffice-qa/
Re: [Libreoffice-qa] writing unit test for calc functions
Hi Markus Personally I'm very interested in using Calc as a replacement for Excel. The spreadsheets I work with commonly have more than 65k lines so using the old Excel 2003 (included in my work desktop) is useless but Calc is still not stable/efficient enough. One of the main problems with Excel and Calc is compatibility. ODS is out of question (because MS manages to ruin all ODS files) and XLSX has lots of issues in round-trip work (particularly in Pivot Tables) Regarding your CSV only approach, Calc does have the option to Save formulae instead of values (and curiously opening these files in Excel 2010 works!) but there are probably limitations (I just found out that array formulas are not interpreted correctly on both programs) Since it is not in Microsoft's interest to be 100% compatible with ODS, I think that your spreadsheet approach should include both ODS and XLSX. So, if manual testing your files leads to greater compatibility you can count me in. Regards, Pedro -- View this message in context: http://nabble.documentfoundation.org/Libreoffice-qa-writing-unit-test-for-calc-functions-tp4181287p4181297.html Sent from the QA mailing list archive at Nabble.com. ___ List Name: Libreoffice-qa mailing list Mail address: Libreoffice-qa@lists.freedesktop.org Change settings: https://lists.freedesktop.org/mailman/listinfo/libreoffice-qa Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/ Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette List archive: http://lists.freedesktop.org/archives/libreoffice-qa/
[Libreoffice-qa] writing unit test for calc functions
Hey guys, so once again I was bit by our missing systematic unit tests for all the formula functions. I'm wondering what we can do to make external people interested in adding tests for all functions. Right now I think that my csv based approach is too complicated (how to handle precision issues, ...) and I wonder if it makes sense to switch from such an approach to a pure spreadsheet approach. So for each function creating a spreadsheet with a summary sheet that contains one cell with a true false return value. This would mean that writing the test is purely creating a spreadsheet and all the error checking happens inside of the spreadsheet. The test would only open the document, force a hard recalc and check that one cell. Basically I'd like to have for each spreadsheet function a document that checks all corner cases and can be used for manual testing as well as for automated testing. Does that make sense? @Winfried: You are writing most of the spreadsheet functions right now. Would that be something that would help you add tests for all new functions? You could add your test cases directly in the spreadsheet and use it for manual checking as well as later for an automated test. @Eike: You are fixing most of the corner cases in the interpreter. Do you think that this is better than our current ucalc based approach for bug fixing? @QA team: Does this sound like something that would make it easier for you to add tests for spreadsheet functions? I'm also open to alternative ideas or people telling me that it is a stupid idea. Regards, Markus ___ List Name: Libreoffice-qa mailing list Mail address: Libreoffice-qa@lists.freedesktop.org Change settings: https://lists.freedesktop.org/mailman/listinfo/libreoffice-qa Problems? http://www.libreoffice.org/get-help/mailing-lists/how-to-unsubscribe/ Posting guidelines + more: http://wiki.documentfoundation.org/Netiquette List archive: http://lists.freedesktop.org/archives/libreoffice-qa/
