Re: 4.1.3_release_blocker denied: [Issue 126622] Base 4.1.2 does not open Tables and Queries in Mac OSX
Is this approved or not approved as a release blocker??? I thought there was a fix for this. It is essential that this is fixed for Mac users. On 2016-10-04, 4:20 PM bugzi...@apache.org wrote: Patricia Shanahanhas denied 4.1.3_release_blocker: Issue 126622: Base 4.1.2 does not open Tables and Queries in Mac OSX https://bz.apache.org/ooo/show_bug.cgi?id=126622 __ -- _ Larry I. Gusaas Moose Jaw, Saskatchewan Canada Website: http://larry-gusaas.com "An artist is never ahead of his time but most people are far behind theirs." - Edgard Varese - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Ready to Vote on 4.1.3?
Please post only objections to calling a vote on 4.1.3 tomorrow. I will treat silence as consent. As far as I can tell, we are ready to go. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Re: In regards to Open Office
Hagar Delest schreef op 04-10-2016 22:04: My fear is: if AOO exports in OOXML (as LibO does), what will happen to ODF? Most users would just use OOXML since it would be compatible with AOO and MS Office. It may lead to frustration because of the glitches from the conversions. OTOH, it may attract new users. Personally: I hardly ever do things because of 'rational reasons' of that kind. I dislike the very idea of OOXML to begin with and as a developer I just lost interest in developing for MS Windows when I was about 18. My favourite format is ODF because I do not use MS Office anyway (haven't used it since about 2000) and "odt" also looks nice in a file browser. But I'm at pains because for me both OpenOffice and LibreOffice are insufficient in terms of quality and robustness and I have started writing stuff in Google Docs because (a) it doesn't crash and (b) it doesn't throw away my text. And for me a vital issue is the poor undo functionality in both programs. Every other program out there has 1000x superior undo functionality as compared to LibreOffice and OpenOffice. The smallest text box in some Browser has better undo functionality than what we have here. And I cannot live without that, because I often have to redo some parts of my text and I cannot constantly save everything for fear I am making a mistake. Hence, today I write in a browser. At least on Linux. On Windows I have options (most notably just Wordpad). - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: In regards to Open Office
Hagar Delest schreef op 03-10-2016 22:57: Even if it came from a previous format, the goal was to make a documented format to allow compatibility with other applications. So not designed from scratch, agreed, but changes made for interoperability. That's how I understand the target of ODF. I just feel that even though people may say they are doing so for "good reasons" in the end you will have found that they just tried to feed their own table. And were doing so for their own reasons. If you are a smaller party and you want people to cooperate with you then it helps a lot if you can show that non-interoperation has no good reasons for it; it is to take those arguments away from critics: now you have no reason to not cooperate. It just stems from the political perspective of someone who finds himself in that position. It's what /anyone/ would do from that position. That doesn't make it better or morally superior; it is just a good strategy to take when you want to be the one they should take up for consideration. So I am just saying it was done for their own reasons and not for 'altruistic' reasons of that kind; many people may say so, but in the end it was just self-interest (and there is not really anything wrong with it and I guess that is the whole point of that). These advocates proclaim moral superiority by pretending to be altruistic and then condemning those who are not the same. But in the end we are all the same and we do things for our own reasons, and the open source advocates do so also. It was /not/ done for altruistic reasons and therefore we are the same as some company who is also not doing it for altruistic reasons. This façade that people are doing things for different reasons than what they are actually doing them for, is what creates the issues. This creates the façade of moral superiority when it is not so; we are all just human, after all. And the whole point of that is that it is okay to do things for your own reasons. You don't first have to "prove" that you are doing something for good reasons before you can go and do it. There is no enemy. Agreed. But Users should be aware of the rules. I must say I found myself in a similar situation when the Opera M2 (mail client) started corrupting my data after I had accidentally started the program twice (at the same time). My email from that period was mostly lost; for how could I ever recover this. So I am not unsympathetic to wanting to be in a place where you can be sure your data is safe. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Testing 4.1.3
I have done some basic testing for the following environments and language versions: Windows 7 en_GB Windows 8.1 en_GB, fr Windows 10 en_GB Ubuntu 16.04 en_GB - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: In regards to Open Office
Jörg Schmidt schreef op 03-10-2016 12:14: From: Hagar Delest [mailto:hagar.del...@laposte.net] I think that ODF was designed to be a fully open standard to give the users back the property of their own data. No, that's not correct. ODF was written this it was compatible with the capabilities of the program OOo. This is a purely technical issue, and does not mean the ODF would therefore not designed open. But [MS Office] OOXML is not what we could label a real open format. There are parts that still refer to proprietary bits. fud, or show me exactly what parts you mean e.g. the old binary formats has MS disclosed, see: https://msdn.microsoft.com/de-de/library/office/cc313105(v=office.14).aspx https://msdn.microsoft.com/de-de/library/office/gg615407(v=office.14).aspx The first version of that document was from 2008. So that's been 8 years now and counting. Coincides a bit with my feelings about it. Microsoft was only fiercely competiting when it was young, mostly also /because/ it was young (and arrogant, hostile). Typically it is the mindset of a young adolescent and I know Bill Gates also had an attitude like that. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: In regards to Open Office
Hagar Delest schreef op 03-10-2016 0:27: In fact, I came to OOo in 2006 because I used to use MS Word to compile data and one day a file got corrupted for an unknown reason. I discovered that there was no way to recover the file because it was proprietary. I think that at that time the .doc format was not disclosed yet (but I may be wrong). What is sure is that I could not get my file back. So I searched the net and found OOo and ODF. And I adopted OOo because of the file format (I had already tried OOo 1 but did not like it). If OOo had not gained popularity, I'm not sure MS would have created something similar with their OOXML. And if the vendor lock-in policy is less an issue now, OOo and ODF may be for something. What it would be if they hadn't been there? Just to precise something: I'm not complaining, and I don't say Linux people are the best or whatever. I just say that file format is an issue. I admit that I think MS do not play fair (but that's logical, else, they would certainly lose users). Up to the user to decide what is more important for him. But if the focus is the application and not the file format, then what? Make a free clone of MS Office? You mean to stick with their format? No, if you are creative you will create your own format to suit your needs, but you don't create your new format for external reasons that have nothing to do with creating an application. The format is useless without application, and since application is its only reason to be the focus is always on the application and not on the format, because the format merely serves that other thing. So I'm only saying that ideological reasons are not a good reason to do anything. It has to have a use also. And I only wanted to indicate, as I discover, that no matter how big of a mouth people have, in the end they are only doing stuff because it works for /them/ also. I think that in the end you find that ideological reasons are what people SAY but not what they DO. The whole Linux ecosystem is so similar to the corporate world that I have started seeing them as the same thing. In this case, there is a point supporting OOXML. But it would slightly become a de facto standard (what was .doc, ...). But would never be fully implemented in all the applications due to the references to the proprietary functions. Well you know that is the façade of open source: that there is any kind of guarantee that the thing /would/ be implemented if the thing was entirely non-proprietary. Without proprietariness there is usually not much of a reason to do anything. I am just saying that open source nature gives no guarantees at all even though they are often projected (but never realized). To go back to GIMP, it is a disorganized whole, there is no organisation almost to its parts. This lack of organisation (both in the developers and the product itself) creates a disorganized picture that is fragmented. Without leadership you cannot do anything. A visionary needs to charge forward himself (or herself, perhaps) and not wait for what others do, but Linux developing is 80% waiting for what others think and do. Before you can do anything. Some people call this "design by committee". It is the death (and dearth) of creativity. I thought back in the day that OpenOffice (certainly under Sun) was an inspired project and it certainly was, I believe. We see today that people like mr. Hamilton would probably not survive the bitterness and alienation and ghastliness of the rock-steady but hostile approach the LibreOffice developers have. There are, in those communities, no elder people that have a bit of wisdom to go with: it is all youth and normally youth that doesn't know much. I am very grateful people such as himself are here. The only older people in such communities are new weds (to the system) that actually know a lot less than the young ones. It is youth and arrogance of youth for the most part, what I see. In communities like what LibreOffice is today... But as mr. Hamilton just said; ODF is not fully implemented in anything. Proprietariness doesn't necessarily mean closed source; it means it is controlled by a single party (for example). It means that single party does not have to suffer design by committee issues. That can also mean people are less likely to adopt it (what you do) but this is a balance by how much you want to cooperate with people and how much you want to plow ahead yourself. Waiting too much for the approval of others creates a dead product. And the only thing LibreOffice is doing is they are improving the /technical/ nature of the code (they are all technical people) and they only promise (better code checking tools and the like) stuff that doesn't mean a better user experience for users, it only means a better user experience for developers. Maybe then after a while they start to focus on the good stuff but thus far they have only done
Re: Testing 4.1.3 - release notes
Could people working on the bugzilla entries check the status and update as appropriate? Note that I did not include the Mac Tables and Queries issue, https://bz.apache.org/ooo/show_bug.cgi?id=126622, in the release notes because I did not see enough confirmation that it is fixed. I have now denied it 4.1.3 release blocker, but added a request for 4.1.4. On 10/4/2016 2:47 PM, Marcus wrote: Am 10/04/2016 10:31 PM, schrieb Patricia Shanahan: https://cwiki.apache.org/confluence/display/OOOUSERS/AOO+4.1.3+Release+Notes ah, thanks. I've updated some text (e.g., separated the sec issue into an own paragraph) and updated the link for the BZ issue list as it still points to 4.0.1. I'm a bit worried about the many issues that are not fixed, resolved, closed [1]. But besides this it's OK. So, we are also done here. [1] https://bz.apache.org/ooo/buglist.cgi?f1=flagtypes.name=equals=Importance_format=advanced=4.1.3_release_blocker%2B Marcus On 10/4/2016 1:29 PM, Marcus wrote: OK, binary and source code testing is on a good way. What about the release notes. Are these somewhere reachable/readable already? - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
4.1.4_release_blocker requested: [Issue 126622] Base 4.1.2 does not open Tables and Queries in Mac OSX
Patricia Shanahanhas asked for 4.1.4_release_blocker: Issue 126622: Base 4.1.2 does not open Tables and Queries in Mac OSX https://bz.apache.org/ooo/show_bug.cgi?id=126622 - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
4.1.3_release_blocker denied: [Issue 126622] Base 4.1.2 does not open Tables and Queries in Mac OSX
Patricia Shanahanhas denied 4.1.3_release_blocker: Issue 126622: Base 4.1.2 does not open Tables and Queries in Mac OSX https://bz.apache.org/ooo/show_bug.cgi?id=126622 - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Question about spell check
On Tue, 4 Oct 2016 13:17:49 -0700 Patricia Shanahanwrote: > > > On 10/4/2016 1:15 PM, Marcus wrote: > > Am 10/04/2016 09:44 PM, schrieb Patricia Shanahan: > >> I downloaded and installed > >> Apache_OpenOffice_4.1.3_Win_x86_install_en-GB.exe. To my distress, when > >> I first started it, it accepted "color" (US spelling) and rejected > >> "colour" (English spelling). > >> > >> I thought perhaps it picked up my previous use of en-US from my profile, > >> but I got the same effect installing on a new Windows 10 computer that > >> has never run any OpenOffice before. > >> > >> It is using English, not American, spellings in the user interface, but > >> I would have expected documents to also default to the installation > >> language. > >> > >> What language is supposed to be used, by default, for user documents? > > > > it's the language from the underlying system. > > > > The same values are set when it comes to other formattings. E.g., you > > will see "US-Dollar" for the currency and not "British Pound". > > > > However, the UI interface is the language from the installation file. > > > > *IMHO* it's OK that not all settings are set to the language of the > > installation file. However, for the default spell check it's > > discuss-worthy. > > I'll check that it is not a regression. If it isn't, we can put it aside > for now. It does seem strange that I need to do more to get en-GB > spelling after installing an en-GB AOO. > On the occasion a User reports this happening we recommend deleting or renaming the OO user profile so it generates a fresh profile on next startup. -- Rory O'Farrell - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Testing 4.1.3 - source builds
Patricia Shanahan wrote: I have built and run from the zip. I have also decompressed and extracted each of the tarballs, and used "diff -r" to confirm they are each identical to the zip. I do plan to do the signature and hash checks for each of the three files. You may want to add your own signature to the .asc files, concatenating it as Dennis suggested. As release documentation explains, this can also be done at voting time, but it's good to keep files unchanged during the vote. If you do so, just remember to use $ svn propset svn:mime-type text/plain *.asc *.md5 *.sha256 (in your case, *.asc will actually be enough) before commit to address the binary vs text issue noted by Marcus. Regards, Andrea. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: 4.1.3 building
Jim Jagielski wrote: Once built and tested, how does one upload? I'm assuming the prepare-download-tree.sh in aoo-devtools?? All builds have already been uploaded by Ariel for this RC. No need to upload anything else. I did upload a full set of Linux-64 builds to http://home.apache.org/~pescetti/openoffice-4.1.3-dev-r1761989/ but due to missing communication (a non-delivered notification to this list) we are using the ones by Ariel for Linux-64 too. prepare-download-tree.sh is just a helper that will rearrange your builds into the right layout. It works for Linux-64, needs minimal tweaks for other systems. Builds (not needed now, since Ariel has already built all systems) must indeed be uploaded through svn commit to https://dist.apache.org/repos/dist/dev/openoffice/4.1.3-rc1/binaries/ which is quite slow; so Ariel committed one binary at a time. I'm not sure how it would work with svn locking if we tried to commit multiple builds to the same dir simultaneously; possibly it will just work as expected, no idea. Regards, Andrea. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Question about spell check
On 10/4/2016 1:15 PM, Marcus wrote: Am 10/04/2016 09:44 PM, schrieb Patricia Shanahan: I downloaded and installed Apache_OpenOffice_4.1.3_Win_x86_install_en-GB.exe. To my distress, when I first started it, it accepted "color" (US spelling) and rejected "colour" (English spelling). I thought perhaps it picked up my previous use of en-US from my profile, but I got the same effect installing on a new Windows 10 computer that has never run any OpenOffice before. It is using English, not American, spellings in the user interface, but I would have expected documents to also default to the installation language. What language is supposed to be used, by default, for user documents? it's the language from the underlying system. The same values are set when it comes to other formattings. E.g., you will see "US-Dollar" for the currency and not "British Pound". However, the UI interface is the language from the installation file. *IMHO* it's OK that not all settings are set to the language of the installation file. However, for the default spell check it's discuss-worthy. I'll check that it is not a regression. If it isn't, we can put it aside for now. It does seem strange that I need to do more to get en-GB spelling after installing an en-GB AOO. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Question about spell check
Am 10/04/2016 09:44 PM, schrieb Patricia Shanahan: I downloaded and installed Apache_OpenOffice_4.1.3_Win_x86_install_en-GB.exe. To my distress, when I first started it, it accepted "color" (US spelling) and rejected "colour" (English spelling). I thought perhaps it picked up my previous use of en-US from my profile, but I got the same effect installing on a new Windows 10 computer that has never run any OpenOffice before. It is using English, not American, spellings in the user interface, but I would have expected documents to also default to the installation language. What language is supposed to be used, by default, for user documents? it's the language from the underlying system. The same values are set when it comes to other formattings. E.g., you will see "US-Dollar" for the currency and not "British Pound". However, the UI interface is the language from the installation file. *IMHO* it's OK that not all settings are set to the language of the installation file. However, for the default spell check it's discuss-worthy. Marcus - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Re: In regards to Open Office
Le 04/10/2016 à 06:50, Peter Kovacs a écrit : ODF has a better trancperency then OOXML. But beeing open we need to embrace and value both formats. This is the key point for AOO now IMHO. I think there is a point having import filters to give users a way to open the files. Then the natural thing would be to save in ODF. That's what was done with OOXML: AOO can import but the code to export in OOXML is not activated. it does exist somewhere but it has never been implemented by design. My fear is: if AOO exports in OOXML (as LibO does), what will happen to ODF? Most users would just use OOXML since it would be compatible with AOO and MS Office. It may lead to frustration because of the glitches from the conversions. OTOH, it may attract new users. Note: it may say that bugs like 1900 leap year bug in OOXML are accepted (except if it has been fixed since). If there was at least the possibility to store the information linked to the AOO features in OOXML without triggering any "conversion" operation, it may be a 3rd way. Does OOXML allow that? I doubt it but perhaps there is room in parts of the format that allow some kind of customization. It would be sad however to lose the human readable structure of ODF (quite useful in case a file is corrupted or when you need to inspect a file in detail). Hagar - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Testing 4.1.3 - source builds
Am 10/04/2016 06:39 PM, schrieb Patricia Shanahan: On some of my Windows builds, I get a failure, but doing a new "build --all", without cleaning, works. That may be worth trying while you are waiting for more expert advice. I think there may be problems in whatever is supposed to be enforcing dependency order, so that a module gets built too soon, while things on which it depends have not all been built. On the comparisons between trees, "diff -r A B" does not take that long and gives full confirmation that A and B are paths to directory trees with the same files and identical file content. I've done a "diff -r AOO413 and aoo-4.1.3" and there I can that the problem is an old "friend". It's the "fmgridif.cxx" file that stumbles over the gcc compiler bug about optimization [1]. In my checked-out SVN files I've worked around that with a modified makefile (thanks to Don). Note to myself: Look closer to the log files. Sorry for the noise. ;-( So finally, the source package as ZIP file is fine and I can get a 4.1.3 release out of it with the release options. [1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65009 Marcus On 10/4/2016 9:11 AM, Marcus wrote: Am 04.10.2016 um 00:30 schrieb Marcus: Am 10/03/2016 11:26 PM, schrieb Patricia Shanahan: On 10/3/2016 2:02 PM, Andrea Pescetti wrote: Marcus wrote: wow, *all signed source code packages* ? I assume that this does not literally mean that you must test the .bz2, the .gz and the .zip. They are equivalent. This sentence is for when a project makes a release composed of different parts. For the record, trunk is already set to avoid duplication of packages, but AOO413 still uses the old convention of 3 source packages. (If it helps, I've used the .bz2 for my tests!). For my testing, I'm assuming that it is enough to be sure a package is identical to one I've tested. In particular, the .bz2 and .gz decompress to the same .tar file, so I don't even plan to extract one of the tar files for further checks. ah, great hint. I've uncompressed all 3 files, diff'ed the .tar.bz2 and .tar.gz files, and finally uncompressed all files until the actual dirs/files. All 3 dirs had the same total file size of 1,541,414,704 bytes. This has to be enough when it comes to "you have to check all source files". Tomorrow I'll build the release from a package file. I've uncompressed the ZIP file and started a clean build. Unfortunately, I get the following error: [...] /share/linux2/aoo-4.1.3/main/solver/413/unxlngx6.pro/workdir/CxxObject/svx/source/fmcomp/fmgridif.o: In function `FmXGridControl::createPeer(com::sun::star::uno::Reference const&, com::sun::star::uno::Reference const&)': fmgridif.cxx:(.text+0x68b2): undefined reference to `non-virtual thunk to WindowListenerMultiplexer::acquire()' /usr/bin/ld: /share/linux2/aoo-4.1.3/main/solver/413/unxlngx6.pro/workdir/CxxObject/svx/source/fmcomp/fmgridif.o: relocation R_X86_64_PC32 against undefined symbol `_ZThn48_N25WindowListenerMultiplexer7acquireEv' can not be used when making a shared object; recompile with -fPIC /usr/bin/ld: final link failed: Bad value collect2: error: ld returned 1 exit status /share/linux2/aoo-4.1.3/main/solenv/gbuild/LinkTarget.mk:259: recipe for target '/share/linux2/aoo-4.1.3/main/solver/413/unxlngx6.pro/workdir/LinkTarget/Library/libsvxcore.so' failed make: *** [/share/linux2/aoo-4.1.3/main/solver/413/unxlngx6.pro/workdir/LinkTarget/Library/libsvxcore.so] Error 1 dmake: Error code 2, while making 'all' 1 module(s): svx need(s) to be rebuilt Reason(s): ERROR: error 65280 occurred while making /share/linux2/aoo-4.1.3/main/svx/prj When you have fixed the errors in that module you can resume the build by running: build --from svx Marcus - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: 4.1.3 building
> On Sep 27, 2016, at 11:03 PM, Ariel Constenla-Haile> wrote: > > > IMO providing different builds for same arch will make it difficult to > QA, so I will go on with the Linux 64-bit builds. If someone wants to > build on CentOS 5 32 bit and upload the binaries, please tell and do so. > Unless someone else picks this up (or has picked it up), let me know and I'll start a build. Still need insights on the upload process on how to get these upstream (unless it's simply an svn copy to https://dist.apache.org/repos/dist/dev/openoffice/4.1.3-rc1/binaries/ PS: I'll be using my ASF signing key. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Question about spell check
I downloaded and installed Apache_OpenOffice_4.1.3_Win_x86_install_en-GB.exe. To my distress, when I first started it, it accepted "color" (US spelling) and rejected "colour" (English spelling). I thought perhaps it picked up my previous use of en-US from my profile, but I got the same effect installing on a new Windows 10 computer that has never run any OpenOffice before. It is using English, not American, spellings in the user interface, but I would have expected documents to also default to the installation language. What language is supposed to be used, by default, for user documents? - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: 4.1.3 building
Once built and tested, how does one upload? I'm assuming the prepare-download-tree.sh in aoo-devtools?? - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Testing 4.1.3 - source builds
On 10/4/2016 4:24 AM, Jim Jagielski wrote: On Oct 3, 2016, at 3:49 PM, Patricia Shanahanwrote: On 10/3/2016 12:45 PM, Marcus wrote: Am 10/03/2016 09:40 PM, schrieb Patricia Shanahan: Testing seems to be going well, but there is a very specific requirement for a release. A PMC member, to cast a binding +1 vote approving a relese, needs to have built the software from source and tested it on a machine under the PMC member's control. See http://www.apache.org/legal/release-policy.html#release-approval PMC members please indicate when they have done that test, to help me decide when to start a vote. I've build today that branch with release options. Is this sufficient or do I need to build from the [zip|gz|bzip] file? I believe it does have to be from the zip etc. but I am not sure. The actual wording is: "Before casting +1 binding votes, individuals are REQUIRED to download all signed source code packages onto their own hardware, verify that they meet all requirements of ASF policy on releases as described below, validate all cryptographic signatures, compile as provided, and test the result on their own platform." The release is the tarball/zip itself and not the "tag". So it (the build) needs to be from the zip/tarball. I have built and run from the zip. I have also decompressed and extracted each of the tarballs, and used "diff -r" to confirm they are each identical to the zip. I do plan to do the signature and hash checks for each of the three files. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Testing 4.1.3
Sorry for the delay: Building OSX as we speak. A build the week-before-last had no regressions. > On Sep 25, 2016, at 10:33 AM, Patricia Shanahanwrote: > > I suggest that people start downloading and testing 4.1.3 as soon as there > are binaries they can run. I can't start the formal vote period until we have > a complete release candidate. > > - > To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org > For additional commands, e-mail: dev-h...@openoffice.apache.org > - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Testing 4.1.3 - source builds
On some of my Windows builds, I get a failure, but doing a new "build --all", without cleaning, works. That may be worth trying while you are waiting for more expert advice. I think there may be problems in whatever is supposed to be enforcing dependency order, so that a module gets built too soon, while things on which it depends have not all been built. On the comparisons between trees, "diff -r A B" does not take that long and gives full confirmation that A and B are paths to directory trees with the same files and identical file content. On 10/4/2016 9:11 AM, Marcus wrote: Am 04.10.2016 um 00:30 schrieb Marcus: Am 10/03/2016 11:26 PM, schrieb Patricia Shanahan: On 10/3/2016 2:02 PM, Andrea Pescetti wrote: Marcus wrote: wow, *all signed source code packages* ? I assume that this does not literally mean that you must test the .bz2, the .gz and the .zip. They are equivalent. This sentence is for when a project makes a release composed of different parts. For the record, trunk is already set to avoid duplication of packages, but AOO413 still uses the old convention of 3 source packages. (If it helps, I've used the .bz2 for my tests!). For my testing, I'm assuming that it is enough to be sure a package is identical to one I've tested. In particular, the .bz2 and .gz decompress to the same .tar file, so I don't even plan to extract one of the tar files for further checks. ah, great hint. I've uncompressed all 3 files, diff'ed the .tar.bz2 and .tar.gz files, and finally uncompressed all files until the actual dirs/files. All 3 dirs had the same total file size of 1,541,414,704 bytes. This has to be enough when it comes to "you have to check all source files". Tomorrow I'll build the release from a package file. I've uncompressed the ZIP file and started a clean build. Unfortunately, I get the following error: [...] /share/linux2/aoo-4.1.3/main/solver/413/unxlngx6.pro/workdir/CxxObject/svx/source/fmcomp/fmgridif.o: In function `FmXGridControl::createPeer(com::sun::star::uno::Reference const&, com::sun::star::uno::Reference const&)': fmgridif.cxx:(.text+0x68b2): undefined reference to `non-virtual thunk to WindowListenerMultiplexer::acquire()' /usr/bin/ld: /share/linux2/aoo-4.1.3/main/solver/413/unxlngx6.pro/workdir/CxxObject/svx/source/fmcomp/fmgridif.o: relocation R_X86_64_PC32 against undefined symbol `_ZThn48_N25WindowListenerMultiplexer7acquireEv' can not be used when making a shared object; recompile with -fPIC /usr/bin/ld: final link failed: Bad value collect2: error: ld returned 1 exit status /share/linux2/aoo-4.1.3/main/solenv/gbuild/LinkTarget.mk:259: recipe for target '/share/linux2/aoo-4.1.3/main/solver/413/unxlngx6.pro/workdir/LinkTarget/Library/libsvxcore.so' failed make: *** [/share/linux2/aoo-4.1.3/main/solver/413/unxlngx6.pro/workdir/LinkTarget/Library/libsvxcore.so] Error 1 dmake: Error code 2, while making 'all' 1 module(s): svx need(s) to be rebuilt Reason(s): ERROR: error 65280 occurred while making /share/linux2/aoo-4.1.3/main/svx/prj When you have fixed the errors in that module you can resume the build by running: build --from svx Marcus - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Testing 4.1.3 - source builds
Am 10/04/2016 10:04 AM, schrieb Andrea Pescetti: Marcus wrote: @Andrea: Can you please check the "apache-openoffice-4.1.3-r1761381-src.tar.gz.sha256" file? It's in binary mode and not useable for checksum comparsion. It can be used if you download it. sorry, no. That's the reason why I posted this. ;-) > But I've now forced all checksum files to be treated as text, which should allow you to click on the file names in https://dist.apache.org/repos/dist/dev/openoffice/4.1.3-rc1/source/ and see them displayed in browser. Thanks for updating the mimetypes. Now it's working. Marcus - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Testing 4.1.3 - source builds
> On Oct 3, 2016, at 3:49 PM, Patricia Shanahanwrote: > > > On 10/3/2016 12:45 PM, Marcus wrote: >> Am 10/03/2016 09:40 PM, schrieb Patricia Shanahan: >>> Testing seems to be going well, but there is a very specific requirement >>> for a release. >>> >>> A PMC member, to cast a binding +1 vote approving a relese, needs to >>> have built the software from source and tested it on a machine under the >>> PMC member's control. See >>> http://www.apache.org/legal/release-policy.html#release-approval >>> >>> PMC members please indicate when they have done that test, to help me >>> decide when to start a vote. >> >> I've build today that branch with release options. Is this sufficient or >> do I need to build from the [zip|gz|bzip] file? > > I believe it does have to be from the zip etc. but I am not sure. The actual > wording is: > > "Before casting +1 binding votes, individuals are REQUIRED to download all > signed source code packages onto their own hardware, verify that they meet > all requirements of ASF policy on releases as described below, validate all > cryptographic signatures, compile as provided, and test the result on their > own platform." > The release is the tarball/zip itself and not the "tag". So it (the build) needs to be from the zip/tarball. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Testing 4.1.3 - source builds
Marcus wrote: @Andrea: Can you please check the "apache-openoffice-4.1.3-r1761381-src.tar.gz.sha256" file? It's in binary mode and not useable for checksum comparsion. It can be used if you download it. But I've now forced all checksum files to be treated as text, which should allow you to click on the file names in https://dist.apache.org/repos/dist/dev/openoffice/4.1.3-rc1/source/ and see them displayed in browser. Regards, Andrea. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Testing 4.1.3 - source builds
Testing the reference builds is indeed extremely important, and should be most of the testing. The significance of the builds from source is that a PMC member can only cast a binding +1 vote if they have done one, and we need at least three binding +1 votes to release. They also need to have a general opinion that the release should go out, and all the non-binding votes and testing reports may influence that. On 10/3/2016 10:45 PM, Mechtilde wrote: Hello, for my understanding, beside doing a good build it is necessary to have and totest defined reference builds. The way I see it, it is not easy to do a good build if you didn't have enough practice doing it. I didn't myself any C/C++ build before. So IHMO I will waste time to improve my build environment instead of testing a reference build My results of testing belong to the reference builds published as RC1 at dist.apache.org. Apache OpenOffice is a project with a wide user base, who only use the binaries. So it is important to release well defined and tested binaries. Otherwise support becomes hell. Kind regards Am 04.10.2016 um 01:58 schrieb Patricia Shanahan: On 10/3/2016 3:30 PM, Marcus wrote: Am 10/03/2016 11:26 PM, schrieb Patricia Shanahan: On 10/3/2016 2:02 PM, Andrea Pescetti wrote: Marcus wrote: - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org Mechtilde Stehmann -- ## Apache OpenOffice.org ## Freie Office Suite für Linux, MacOSX, Windows ## Debian ## Loook, calender-exchange-provider, libreoffice-canzeley-client ## PGP encryption welcome ## Key-ID 0x141AAD7F - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: Infra notification?
Patricia Shanahan wrote: According to http://www.apache.org/legal/release-policy.html#heads-up, infra needs to be notified in advance of releases of more than 1GB. This used to be important to avoid flooding the mirrors. The mirrors that had disk space issues are now in a separate list which does not mirror OpenOffice. So yes, this still applies (and it is unrelated to hosting the binaries on SourceForge, as we keep a copy at Apache too; just, Apache has the disk space but not the bandwidth); and it is enough to send a quick mail to Infra for acknowledgment, exactly as you have done in the meantime. Regards, Andrea. - To unsubscribe, e-mail: dev-unsubscr...@openoffice.apache.org For additional commands, e-mail: dev-h...@openoffice.apache.org
Re: In regards to Open Office
On Mon, Oct 3, 2016 at 11:53 AM, Jörg Schmidtwrote: > > > and a very little bit is my opinion also: > MS is one of our "Platinum sponsors" and it is not a good style blindly to > grumble about MS > > Maybe, but sponsoring the ASF also shouldn't undermine contributors' free speech. > > > Greetings, > Jörg > > Damjan