I have been able to edit a spreadsheet and use the menus just fine with (almost) current 3.28. A big problem I have recently noticed is that I am no longer able to Upload new docs. The windows work and the file selector opens, but I get a DSI when the file is chosen, and the filename is shown corrupted and will not upload. I know this used to work waaay back a year or so ago, since that is how I uploaded a couple docs there in the first place.

Can someone else confirm this Upload bug? I will make some crashlogs and add to bugzilla if it's not just me.


Scott Woodall

PS: I am still running 4.1 without upd 1 or 2 because of hyp. website registration failures

If you are someone who has had an AmigaOne for a long time and especially if you leave it on all the time like I do, It is time to replace your fans!

Most fans are only made to last for a few years, and they will often start to perform worse and worse as they age, which means you will start heating up. If the bearings on your CPU or case fan die, you may not hear it or notice before you have a dead machine. It's getting into spring overheating season in the northern hemisphere, so when you take your case apart for dust removal, spend a few quatloos to replace your important fans and hopefully the A1s will live a while longer.

@MickJT

The 'which' command is only for finding an executable program in the current path, not for what you want here.

To view all of an assign, you can use 'assign libs: exists' or similar. See this and other fine usage information in your sys:documentation/c/assign.doc

Yes it is normal for an assign command with no target to delete the assignment. It is treated as "assign libs: to nothing"

Because assignments are a _list_ of paths to search, when you REMOVE you need to specify which path in the list you want removed.

The SDK startup scripts contain assigns, paths, and env var settings. Some of those are system-wide changes and some are for the shell only. So if you run it in one shell, that shell and others opened from it are the only places it is fully active. If it's done in user-startup, that runs in the startup task before it launches 'LoadWB', which means the Workbench inherits all the changes from it, which also means everything else like shells launched later also inherit it all.


Scott Woodall

@rwo

Quote:
Is that a Freudean slip? You know how expensive gold is right now! :)


@abalabalaban

In addition to monitoring, etc, USB control is actually a way to work around the lack of software poweroff capability in the hardware. Just tell the UPS to power off.


Scott Woodall

@Mrodfr

Partly what we want is a piece of the roadmap -- what do the developers have targetted for development next (if any) and how can we provide input to that prioritization decision. Where are the areas where we need unified support within existing drivers like HID vs where non-core developers could serve buy creating specific function drivers.

If we know what the planned direction and timeline looks like, then there may be room for bounties or other ways we can influence additional development.


Scott Woodall

PS: I would totally buy someone a USB UPS to build a good driver for an 'apcupsd' equivalent.

@tlm

Sounds like maybe just a bad port on the hub. Either way, you have the solution options before you now.


Scott Woodall

Quote:
Sounds like a lower level problem than name servers. Can you ping your router from the SAM? (assuming your other computers can)

What does the output of this command look like on your SAM (substitute your connection name for "lan0"):

There are many options to shownetstatus, maybe ARP, ROUTES, DNS, or RT will help you.

Scott Woodall

@Slayer

Are you just talking about "turning it off" via the DPMS control? On my blanker, Shift-F1 blanks the screen, and Shift-F2 turns the monitor off this way. If you don't want it to blank at other times, you can control the blanker timeouts also. That and the hotkeys are setup in the ScreenBlanker prefs that comes with the OS.

Scott Woodall

@Atheist

That example on the page is incorrect. I don't see anything in the documentation drawers about pattern matching, but this has been mostly the same since OS 1.x

Maybe this online guide will help you.

http://www.titan.co.nz/Amigaak/AA020844e00.htm#PATM


Scott Woodall

I have encountered a new 4.1 problem; can a betatester verify and get in a bug report if needed please.

Start mplayer showing a video, and use the resize drag gadget in the lower-left to resize it narrow so the horizontal size goes to zero width -- it will lock up all graphics. I couldn't find another application that would let you resize the window to 0 width because of some contents, but I suspect it's a layers or intuition problem and not just mplayer. It doesn't happen on 4.0 just 4.1. Anyone else ?

Scott Woodall

@ldituri

Those three *.so files in the OWB directory -- did you copy them to SYS:SObjs also? Can you show us the shell output from `list SOBJS:` ?

Scott Woodall

@GrumpyOldMan

My card sounds identical to yours and mine is 64-bit. I don't know if there is any way to tell directly unless you have the original packaging, and even then, who knows.

Scott Woodall

@Lio

You are correct, and I can duplicate the same problem. I deleted the first partition and a partition at the end of a disk, and it would not let me add the end partition back when there was empty space at the low end. (MTB 52.2)

Scott Woodall

@Lio

I recently setup a disk with 20 partitions. How are you trying to create the new partitions -- are you using only unused space on the HD, or are you shrinking existing partitions to split into multiple?

Scott Woodall

@PEB

Oops, I completely forgot that those pointer defs files override the pointer prefs setting.

Seems like a good opportunity (ahem) to prevent this mistake in the future if the prefs app would TELL you when you use/save that there was a global pointer def in use which would override any changes you made there. Kind of a tease, isn't it? Maybe rename it "PointerPrefsButNotThePNGOnesJustTheOldBitmappedOnesKThx". Same sort of tease with IconEdit (don't try it on a PNG icon, or you'll have to wash your retinas with bleach to remove the afterimage of ugly). Great to have the PNG options, but a shame they didn't get prefs/control integrated better (yet?).

Would adding some simple PNG truecolor gfx edit capabilities into OS4 be enough to integrate it into these prefs? How are the pointer and iconedit bitmap editors written, using some shared system code, or are they both custom fully coded in these functions? Or at least make the prefs able to load and display the PNG types and be able to launch your own configurable app to edit if you have one (ImageFX, etc).

Glad he got his plain ol' pointer back.

Scott Woodall

@MamePPCA1

Tell us what you have tried which didn't work. I don't think the pointer is changed by the themes I have seen.

This works for me:
- open the prefs window
- launch Pointer
- back to workbench, open the Presets drawer in Prefs
- open the Pointers drawer, and "Show/All files"
- drag and drop HiRes-Colored.pre into the Pointer editor
- save
- dance for joy
- remember to set clock ahead 1 hour

Maybe you can skip that last step. ;)


Scott Woodall

@Chris

Your thread seems to be getting a good response and accumulating the data asked for. However, now I am wondering if it's the right information which is needed.

The vast majority are using interrupts ON, and I don't see that those who don't are clustering into groups by card type or other patterns. Should we be asking explicitly Why do those without interrupts enabled have that setting? Is there a specific function they did it to correct, or an application which needs it, or ? Have they tried it ON or is that just how it has been set since they tinkered with it for some other reason? Etc.

Scott Woodall

@Chris

14 0x1002 0x5C61 ATI Technologies Inc RV250 5c61 [Radeon Mobility 9200 M9+]
Card is a "Sapphire 9250 128M" AGP 64-bit mem

INTERRUPT=Yes

ATIRadeon.chip 52.5
A1XE G4


Scott Woodall

@ikir

The wikipedia does not list all permutations of features, speed, memory width, etc. If you are looking for fastest, you don't want to get any card that says what these cards do in the specs list: Quote:Quote:

There are quite a lot of 9200s and 9250s out there without a full 128-bit memory interface. If they don't mention in eBay or on the box, assume it is the slower (cheaper) 64-bit memory version. Many of the auctions are just the text of the capabilities of the RV280 chip -- not the specific card they are selling. If it says something like "Available in both a 128-bit and 64-bit memory data path interface", that's just the GPU spec, so keep looking for more description of the model itself. I haven't ever seen one of the 256Mb versions with a 64-bit interface that I know of.

Edit:
Here is a two year old article (about the time they came out) about 9250s.


Scott Woodall