On 07-11-2014 14:37, Bruce Dubbs wrote: > Fernando de Oliveira wrote: >> On 07-11-2014 13:35, Bruce Dubbs wrote: >>> Ken Moffat wrote: >>>> On Fri, Nov 07, 2014 at 12:53:33AM -0600, Bruce Dubbs wrote: >>>>> I'm trying to build seamonkey-2.30 on an older 686 running a 2 year >>>>> old >>>>> version of LFS. After about 3.5 hours, the build crashed with >>>>> messages >>>>> like: >>>>> >>>>> jsapi.cpp:(.text+0x1db8): undefined reference to `u_init_52' >>>>> >>>>> but there are a tone of these with the suffix _52. >>>>> >>>>> I do have the option set: >>>>> >>>>> ac_add_options --disable-optimize >>>>> >>>>> On a newer build (x86_64), SM builds builds fine. >>>>> >>>>> I think I saw something about this some time ago, but I can't find it >>>>> now. >>>>> >>>>> Does anyone remember the issue and the solution? >>>>> >>>>> -- Bruce >>>> >>>> We had something similar in LO. The _52 is the version of icu4c. >>>> /me searches ... Apparently, I also hit this in firefox 29.0 >>>> (like you, that was on an _old_ system, mine was probably in the 7.0 >>>> to 7.2 range). I'd forgotten that, but google remembers: >>>> https://bugzilla.mozilla.org/show_bug.cgi?id=971663 >>>> >>>> Try using system icu, if you have one. >>> >>> It was late last night, but I had seen that reference. I rebuilt SM >>> with 'ac_add_options --without-intl-api' because I really didn't want to >>> take the time to build icu and I don't need any i18n on that system. >>> This morning it had completed OK but only after 54.6 SBU (about 3.5 >>> hours) on this machine. >> >> I spent some time this morning with this, but couldn't find, always FF >> was in the search. As ĸen and you, I remembered there was a related >> problem. >>> >>> On newer systems is seems to build OK. I wonder if we should mention >>> something about this on the SM page in the book? In any case ICU should >>> probably be mentioned as an optional dependency. >>> >>> Thoughts? >> >> Well, I think SM, FF and TB might have the same dependencies or almost, >> but you all know how weak I am about dependencies. >> >> For reference, FF has: >> >> {{{ >> Recommended >> >> ICU-54.1, libevent-2.0.21, libvpx-1.3.0, NSPR-4.10.7, NSS-3.17.2, >> SQLite-3.8.7.1 and yasm-1.3.0 >> }}} >> >> Notice that some of them appear as optional, in SM. > > I installed SM-2.30 and it works. It picked up my mail and browser > config OK and that is always a concern for me as those are my most > important apps. > > A couple of notes. > > 1. I built without ICU.
In my LFS7.1 host (SVN-20120311), I use exactly the same as for LFS-7.6. All with system ICU > 2. It took a long time to launch the first time, but I suppose that's > OK. It seems slower, e.g. reply to to this message is my first reply > with SM, but that may resolve itself. This is the result of the non-optimization. Worse, it regularly makes something taking 100% of one CPU for a dozen seconds. First launch is slow for the same reason. Latest 33.0.3 that I have just updated builds correctly with enable-optimize and starts much faster. Still have to build and test in this host. Probably, next SM version will be better in all that. > 3. The https://github.com/libical/libical/ site I've been checking > still does not have icons instead of blocks with F0xy in them. That may > be due to no ICU or something else. It probably does not need to be > addressed as it only occurs in fairly old systems. As I said in other thread, it happens in the latest versions of SM and FF (just tested in FF-33.0.3). But it is displayed correctly in Midori. > > 4. I used to have some issue with horizontal breakup when scrolling > some sites like slashdot. That seems to be fixed. I didn't have that. > 5. The book says: > > "This package does not come with a test suite. However, if X is running > it can be launched from the build directory before installing with the > command line: moz-build-dir/mozilla/dist/seamonkey/seamonkey" > > However there is no seamonkey in moz-build-dir/mozilla/dist/ until after > install. > > There are two identical instances in: > ./moz-build-dir/mozilla/dist/bin/seamonkey > ./moz-build-dir/suite/app/seamonkey > > but they do not work. There is an XPCOM problem of some kind. > > After install (over my old version), > moz-build-dir/mozilla/dist/seamonkey/seamonkey did appear, but that's > too late to test. I didn't try a DESTDIR install. > > We should probably just remove the 2nd sentence: "However, if X is Agree. -- []s, Fernando -- http://lists.linuxfromscratch.org/listinfo/blfs-dev FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page