OSX: Tess.init crashes after computer reboot
BlitzMax Forums/Brucey's Modules/OSX: Tess.init crashes after computer reboot
| ||
Hello everyone, hello Brucey, I started using your tesseract module, as I find it simpler to use the inbuilt module instead of a precompiled package. Under MacOS 10.5 it worked fine for me - until I shut down the computer and started it up the next day. Since then the programme crashes during "Tess.init("tessdata","eng")". I tried to dig a bit into it, but I'm getting quite confused. It seems to me, that there might be a problem with the installed dictionaries - but I'm not sure. I also tried to install it under Linux openSuse 11.3, but I failed to compile it. So I can't give you any reports about the linux version, sorry. Many thanks for any help. Greetings Simon |
| ||
The tesseract module has now been updated (in SVN) to the latest 3.x version, which I presume makes for better character recognition. I think there is also some improved support for layout. It now also requires BaH.Leptonica, an imaging library, also available in SVN. I've tested it on the 3 platforms, and it appears to work as expected. |
| ||
... Last edited 2011 |
| ||
Yay, this is werid. After releasing that the above references to mingw 3.4.5 and I am running a newer version, I decided to reinstall MinGW altogether. Now it goes completely bonkers, with garbled command line output: C:\Code\BlitzMax\bin>bmk makemods -a bah.tesseract Compiling:strtok.cpp Compiling:baseapi.cpp Compiling:tesseractmain.cpp In file includeInd fi le fincrludoed mfro m CC:/Co:d/eC/oBdlei/tBzlMiatxz/Mmaoxd //mboadh/.bmaohd./mloedp/ttoensseica.morda/srcc/atllt.ypems.ho:20d:0,/ s r c / w i frnom 3C:/2Cod/e/BlsittrzMatx/mokod/bah.mod/le.ptocnpipca .mod/src/allhe:a1der:s.h0::23, c : / m i nfrogm Cw:/C/odel/BliitzbMax//mogd/bcah.cmod//temsseiracnt. mgod/wsrc3/ap2i/basea/pi.4cpp:27: c:/mingw/lib/gc.c/mingw32/4.5.50/...0//../.../...//inc.lud.e/s/tdi.o.h:26:.20:/ fa.tal. er/rori: sntddcef.h: No such file or dlireuctodrye /cosmtpilartiionng t.ehrm:i2na4t:e20I:dn ff.ai ael ienrcrlourd:e ds tfdrdoemf .Ch::/ CNood es/uBclhi tfziMlaex /omro dd/ibraeh c.tmoordy/te sercacto.mBomdu/siprcl/icdcu ltiEla/rmfrtcopcrih.:h: 1o2:0f, i l te d e rt o m icfroonmm Cap:/iCtoldeee/B ldi(tz.1Ma)x mCo:d//Code/BlbiatzhM.amxod//tmeossder/acbt.amohd/.srmc/oadpi//ttesesesrascetmra ainc.ctpp.:2m0:o dc://msinrgwc/l/iba/pgcic//mibngaw3s2/e4.a5.p0/i.../.c./p.p./ /i ncBuild Error: failed to compile (1) C:/Code/BlitzMax/mod/bah.mod/tesseract.mod/ src/api/baseapi.cpplBude/stdlib C:\Code\BlitzMax\bin>.h:21:20: fatal error: stddef.h: No such file or directory compilation terminated. |
| ||
Well, that's very cool ! I coded it initially on Mac. I tested it on Windows (and fixed all the broken MinGW stuff) Then I tested it on Linux, which worked first time. Testing consists of rebuild after rebuild until all the bits compile. And then I run the examples, applying other fixes as required. I think I'm using a v4.x MinGW ... I'll give it another clean build and see if anything crops up. Last edited 2011 |
| ||
FYI, I'm building with TDM GCC 4.4.1... which I think is a fairly standard build. Unless of course there are problems building on something older (non 4.x gcc)... ho hum. |
| ||
OK, it took near-divine intervention, but I managed to get MinGW back into whack to where it can successfully compile the module. Did I ever mention that I hate MinGW? :-? |
| ||
(I'm using tdm 4.5.1, btw) |
| ||
By the way: it appears that the garbled text was BMK NG doing something funky. Trying it with the original BRL bmk resulted in normal readable error messages. Everything appears fine now, though -- recompiling all my modules at the moment, and so far I'm not seeing any more error messages or garbled text. |